How does Rasch differ from classical test theory?
Rasch measurement differs from CTT in several appreciable ways. First, CTT is primarily concerned with total scores. Total scores, the number of items answered correctly, serve as the sole indicator of a person’s level of ability or knowledge, and all items are treated as equal contributors to the total score. That is, difficult items are not weighted more highly than easier items when estimating levels of knowledge or ability. Moreover, equal differences in total scores are treated as delineating equal ranges of ability, whether they be from 40 to 50 points, for example, or from 85 to 95 points. In CTT, the primary indicator of test quality is reliability, the overall capacity of the test to define levels of knowledge or ability consistently. Reliability estimates are usually derived by comparing the total scores of one half of the test to the total scores of the other half, either directly, or indirectly using formulas such as Cronbach’s alpha that approximate the mean correlation of
